Obesity indices in the prediction of incident type 2 diabetes mellitus in a long- term cohort study of Hong Kong Chinese

Abstract

Aim: Obesity is closely related to the development of type 2 diabetes mellitus (T2DM). We investigated the usefulness of various obesity indices in the prediction of incident T2DM in Chinese, based on the Hong Kong Cardiovascular Risk Factor Prevalence Study (CRISPS) cohort with up to 17 years of follow-up. Methods: The Hong Kong CRISPS is a population-based prospective cohort study of Chinese men and women, aged 25-74 years, initiated in 1995. Subjects were contacted for regular reassessment visits from 1995 to 2012 (CRISPS 1-4). A total of 1512 subjects with no diabetes at CRISPS1, who returned for follow-up at CRISPS4, were included for analysis. Cumulative incidence of T2DM, based on oral glucose tolerance test or taking anti-diabetic drugs, was ascertained until December 2012. The associations of baseline obesity indices at CRISPS1, including body mass index (BMI), waist circumference (WC) and waist-hip ratio (WHR), with the development of T2DM were evaluated by logistic regression. The optimal cut-offs of these indices in the prediction of incident T2DM were evaluated by receiver operating characteristics curve. Results: A total of 286 subjects (18.9%), aged 43.2 ± 10.9 years at CRISPS1, developed T2DM over 17 years of follow-up. Their baseline BMI, WC and WHR were 24.1±3.53kg/m2, 78.6±9.85cm and 0.83±0.08 respectively. All three indices independently predicted the development of T2DM after 17 years (BMI sex-adjusted odds ratio [OR] 1.25, 95% confidence interval [CI] 1.20-1.30, p <0.001; WC sex-adjusted OR 1.10, 95% CI 1.08-1.11, p <0.001; WHR sex-adjusted OR 1.09, 95% CI 1.07-1.12, p <0.001). The area under the curve (AUC) for BMI, WC and WHR were 73.0 (95% CI 69.7-76.3), 72.8 (95% CI 69.6-76.0) and 71.5 (95% CI 68.4-74.7) respectively, although Delong test did not suggest superiority of any obesity index over the others. The optimal cut-off of BMI was 24.5kg/m2, which yielded a sensitivity of 70.3% and a specificity of 66.5%. Using a BMI cut-off of 23kg/m2, the sensitivity increased to 83.2% with a specificity of 45.5%. Conclusion: Simple obesity indices are useful tools in the prediction of incident T2DM in Chinese. A BMI cut-off of 23kg/m2 provides good sensitivity with reasonable specificity in our population. Our findings are in keeping with the latest recommendations from the American Diabetes Association, not only in using BMI as a prerequisite for screening, but also the BMI cut-off used when testing asymptomatic Asian adults for diabetes.